PubMed1998
Vertebral fractures are a common and important consequence of osteoporosis and are often identified via morphometric analysis of conventional lateral spine radiographs (morphometric radiography or MRX). A new method of performing vertebral morphometry using images acquired on dual-energy X-ray absorptiometry (DXA) scanners (morphometric X-ray absorptiometry or MXA) has recently been developed. In this study, we derive reference data for vertebral heights and height ratios using MXA scans as the data source and compare the results with previously published MRX studies. One thousand and nineteen Caucasian women (mean age 63 years, range 33-86) were recruited. An MXA scan, covering 13 vertebrae from L4 to T4, was acquired for each subject on one of four DXA systems located at three centers in the U.K. Analysis of variance found statistically significant but relatively small differences among centers, machines, and scan modes, and therefore data were pooled for reference range calculations. Three vertebral heights (anterior, mid, and posterior) were measured and four ratios (wedge, mid-wedge, and two crush) calculated. These data sets were trimmed using an iterative algorithm to remove extreme values assumed to represent deformed vertebrae, then mean and SD values were calculated using the remaining data. When the data were split by age, a small but statistically significant decrease in vertebral height between the sixth and eighth decades was found, but this was not replicated for the vertebral height ratios. Marked differences were observed between MXA data and MRX, but were comparable to those between different MRX studies. These may result from differences in image quality and point placement protocols, population differences, differences in radiographic technique, and differences in the derivation of a group of "normal" vertebrae. This study suggests that reference data of vertebral dimensions should be specific to the technique which uses those data as a reference, i.e., MXA.